Re: Portability warnings: split into non-GNU and non-portable?

[Reuben Thomas]

On Sat, 27 Jan 2007, Ralf Wildenhues wrote:

> * Reuben Thomas wrote on Sat, Jan 27, 2007 at 08:25:41PM CET:
> > On Sat, 27 Jan 2007, Ralf Wildenhues wrote:
> >
> > > Why not, I guess.  Except that automake isn't really able to distinguish
> > > between GNU extensions and a large class of typos or otherwise malformed
> > > makefiles.  Bothered enough to write a patch?  ;-)
> >
> > To clarify what I meant, I meant splitting the non-portable constructs it
> > currently detects into a) GNU extensions and b) other non-portable
> > constructs.
>
> That's how I understood it.  My point is that the current Automake code
> doesn't "detect" GNU extensions.  It cannot distinguish between correct
> GNU make extensions and plain bugs.  Merely not warning about any of
> those would be a regression for some cases: for example, they may not be
> noticed by 'make' on your system, say, because on your system some
> Automake conditional branch won't be expanded.

Can you point me to the bit of code I should look at, please? It sounds like 
automake is just verifying the syntax of the Makefile.am and not actually 
flagging specific constructs. Is that right?

In that case, I understand your concerns.

--
http://rrt.sc3d.org/ | poetry, n.  the art of squeezing blood into a stone